Re: [Gen-art] Gen-ART review of draft-ietf-avtext-client-to-mixer-audio-level-05.txt

[Emil Ivov <[e-mail]>]

Hey Alexey,

On 27 sept. 2011, at 00:24, Alexey Melnikov <[e-mail]> wrote:

> Jonathan Lennox wrote:
> 
>> Hi, Alexey -- thank you for the Gen-ART review.
>> 
> Hi Jonathan,
> 
>> Alexey Melnikov writes:
>> 
>>> Question: are the two encoding of the audio level indication option specified in the document really necessary?
>>>   
>> 
>> Do you mean the one-byte vs. two-byte forms of the header extension (Figure 1 vs. Figure 2)?  These are the two forms of the generic header extensions defined by RFC 5285.
>> 
> I understood that. Does RFC 5285 require that both forms should be allowed?

It doesn't explicitly say so but it It actually does, yes. Here's what it says:

   A stream MUST contain only one-byte or two-byte
   headers: they MUST NOT be mixed within a stream.

Audio level headers can find themselves in streams that also have other, longer extensions, which do require the two-byte header. The above lines mandate that in such cases they all use the two-byte header.

In the same regard, although probably a bit less likely, nothing prevents having another sixteen header extensions in a stream that also has levels. In that case we'd need to switch to two-byte headers in order to be able to fit all the IDs.

Cheers,
Emil

--sent from my mobile

> In general, it would be good to avoid multiple representations of the same thing.
> 
>> The actual payload (one byte containing the V and level bits) is identical in the two cases; the only difference is the container.  We can add some text clarifying this point if you think it would be helpful.
>> 
>>> Nits/editorial comments:
>>> s/relys/relies ???
>>>   
>> Thanks, will fix.
>> 
>